ecological impact of the “erika” oil spill: determination of the

Aquat. Living Resour. 17, 369–377 (2004)c© EDP Sciences, IFREMER, IRD 2004DOI: 10.1051/alr:2004036www.edpsciences.org/alr

AquaticLivingResources

Ecological impact of the “Erika” oil spill: Determinationof the geographic origin of the affected common guillemots

Bernard Cadiou1,a, Léa Riffaut2, Karen D. McCoy3, Jérôme Cabelguen4, Matthieu Fortin4,Guillaume Gélinaud4, Alexandra Le Roch4, Claire Tirard5 and Thierry Boulinier2

1 Bretagne Vivante - SEPNB (Société pour l’Étude et la Protection de la Nature en Bretagne), 186 rue Anatole France, BP 63121,29231 Brest Cedex 3, France

2 Laboratoire d’Écologie, CNRS - UMR 7625, Université Pierre et Marie Curie, 7 quai Saint Bernard, 75005 Paris, France3 Department of Biology, Queen’s University, Kingston, Ontario, Canada4 Bretagne Vivante - SEPNB, (Société pour l’Étude et la Protection de la Nature en Bretagne), Brouel Kerbihan, 56860 Séné, France5 Laboratoire de Parasitologie évolutive, CNRS - UMR 7103, Université Pierre et Marie Curie, 7 quai Saint Bernard, 75005 Paris, France

Received 29 January 2004; Accepted 29 July 2004

Abstract – Between 80 000 and 150 000 marine birds wintering in the Bay of Biscay were killed during the “Erika”oil spill. Three complementary studies were conducted to investigate the geographic origins of these birds. The commonguillemot, Uria aalge, represented more than 80% of the oiled birds and these studies thus focused primarily on thisspecies. Analyses of 184 ring recoveries and biometry of 1851 corpses indicated that guillemots originated from alarge geographic area, including colonies from across the British Isles and the North Sea, along with more northernlocalities. However, the majority of individuals came from colonies located between western Scotland and the CelticSea. The third study, based on a population genetic approach using microsatellite markers (samples from dead oiledbirds and from more than 600 birds caught in 19 breeding colonies), showed little genetic differentiation among north-eastern Atlantic guillemot colonies. This result limits the ability to identify the geographic origins of the birds usingonly DNA samples, but reveals a significant amount of gene flow among colonies. Overall, results indicate the largespatial scale of the oil spill’s impact and underline the usefulness of combining multiple approaches to assess the localand regional effects of such accidents.

Key words: Oil spill / Ring recoveries / Biometrics / Genetics / Seabird / Uria aalge / Atlantic Ocean

Résumé – Détermination de l’origine géographique des guillemots de Troïl victimes de la marée noire du pé-trolier « Erika ». La marée noire de l’« Erika » a décimé plus de 80 000 oiseaux marins hivernant dans le golfe deGascogne, avec une estimation maximale de 150 000 victimes. Trois études complémentaires ont été mises en placepour évaluer l’origine géographique des oiseaux impactés. Le guillemot de Troïl, Uria aalge, représente à lui seul plusde 80 % des oiseaux mazoutés et ces travaux concernent principalement cette espèce. Il s’agit de l’analyse des reprisesd’oiseaux bagués (portant sur 184 guillemots bagués comme poussins et retrouvés échoués), d’une étude biométrique(réalisée sur 1851 cadavres) et d’une étude génétique (réalisée sur des échantillons d’oiseaux échoués et sur plus de600 prélèvements provenant de 19 colonies de reproduction). Les résultats des deux premières études montrent que lesguillemots sont originaires d’une vaste zone géographique, englobant l’ensemble des colonies des îles Britanniques etde mer du Nord ainsi que des colonies plus septentrionales, mais en proportion variable. En effet, la grande majoritédes oiseaux touchés est principalement originaire des colonies situées entre l’ouest de l’Écosse et la mer Celtique.La troisième approche a utilisé des marqueurs génétiques neutres (microsatellites). Elle montre une faible structura-tion génétique des populations, ce qui limite les possibilités d’identifier par cette méthode l’origine géographique desguillemots, mais souligne l’importance des flux de gènes entre colonies à l’échelle de l’Europe du nord. L’ensemble deces résultats montre la grande étendue géographique de l’impact de la marée noire et souligne l’intérêt de combiner desapproches à différentes échelles pour préciser les effets locaux et régionaux d’un tel accident pétrolier.

a Corresponding author:[email protected]

370 B. Cadiou et al.: Aquat. Living Resour. 17, 369–377 (2004)

1 Introduction

On the 12th of December 1999, the tanker Erika broke upin the north of the Bay of Biscay, off the coast of Brittany.The amount of oil released at sea was about 20 000 tonnes andmore than 400 km of the French coast was impacted with var-ious degrees of oil pollution. The first oiled birds were foundin the northern part of the Bay of Biscay between the 14th andthe 18th of December (Cadiou et al. 2003a). From the avail-able data compiled for the national assessment of the seabirdwreck and mortality following the oil spill (“Bilan nationaldes échouages et de la mortalité des oiseaux”; Cadiou et al.2003a), about 74 000 oiled birds (32 000 alive and 42 000 dead,mainly seabirds and seaducks) were recorded ashore along thecoast of the Bay of Biscay. As in the case of other oil spills(Piatt and Lenski 1989; Heubeck et al. 2003), determiningthe geographic origin of these birds is of major importanceif one wants to assess the ecological impact of the oil spill.One species, the common guillemot Uria aalge represented83% of the birds killed. The French breeding population ofthe common guillemot is limited to only a few hundred pairsbreeding in Brittany (Cadiou 2002), but several million pairs ofthis species breed along the coasts of northern Europe (Harrisand Wanless 2004), which underlines the need to identify theorigin of the birds.

Amongst other studies financed by the French governmentwithin the scientific network set up to assess the consequencesof the Erika oil spill, three concerned the determination of thegeographic origin of the seabirds killed after the wreck. Theywere based respectively on analyses of ring recoveries, biomet-rics of birds and genetic data. This paper presents the resultsobtained for the common guillemot.

2 Materials and methods

2.1 Ring recoveries

Ringed birds were collected and recorded by various peo-ple or rehabilitation centres receiving oiled birds. However,due to the difficulty to detect ring on heavily oiled birdsor to the lack of organization, it appeared that about halfof the ringed birds have not been recorded (Cadiou andDehorter 2003). Data were centralized by “Bretagne Vivante”and CRBPO (“Centre de recherches sur la biologie des pop-ulations d’oiseaux”, Museum National d’Histoire Naturelle,Paris) and sent to ringing centres abroad to obtain informationon the date and locality of ringing, and the age of the birds atringing.

2.2 Biometrics

About 2300 corpses of oiled birds collected directly onthe coast or coming from rehabilitation centres were used forthe biometric study. The study sample included 1851 commonguillemots. During external and internal examinations, severalmeasurements were made according to standardized methods(Hope Jones et al. 1982a; Camphuysen 1995). Many species ofbird exhibit phenotypic variation in structural size or plumage

pattern that can be related to their natal area and thus can beused to identify the origin of birds killed during an oil spill(Hope Jones et al. 1982b; Anker-Nilssen et al. 1988; Weir et al.1997). Some body characters can also vary according to age orsex, allowing to infer information on the structure of the af-fected group of birds (sex-ratio, age ratio, sexual maturity). Inthe common guillemot, there is a latitudinal gradient in upperparts colour and wing length, birds originating from northerncolonies being darker and having longer wings (Cramp 1985;Hope Jones 1988). Data on plumage colour and wing lengthof mature birds from the oil spill were thus compared to whatis known about these latitudinal variations. Comparisons weremade according to the sex of individuals as males and femalesdiffer in size. The cloacal bursa (or bursa of Fabricius) is an or-gan well developed in young birds (Møller and Erritzøe 2001).Seemingly, white tips on underwing coverts are present infirst-year guillemots (Camphuysen 1995; Cadiou et al. 2003b).Thus these two variables were considered as surrogates of theage of a bird. The sample included 952 females, 695 malesand 204 unsexed birds and all measurements were not alwaysavailable for all the birds.

2.3 Genetic data

DNA can be extracted from the tissue of birds recov-ered dead or alive following an oil spill and the correspond-ing genetic information could be used to determine the ori-gin of the birds. The objective of the third approach was thusto test whether it was possible to assign individuals to theirbreeding area of origin using an assignment method (Cornuetet al. 1999) and data on the genotypes of individuals fromthe oil spill and from breeding colonies (Riffaut et al., unpub-lished). Such an approach is dependent on the ability to genet-ically characterize sub-groups of individuals. The applicationof such an approach was attempted because of the potentiallygenetically structured populations of seabirds (associated withtheir high philopatry and the description of phenotypically dis-tinct groups) and the recent availability of highly polymorphicmarkers (microsatellites), which appeared to be promising touse in this context (Edwards et al. 2001). The method usesthe allele frequencies of each sampled population to computethe probability of occurrence of the multi-locus genotype ofindividuals within each population.

In order to genotype individuals, DNA was extracted fromblood, growing feathers or muscle tissue (in the case of deadbirds). The study involved two main steps (Riffaut et al., un-published). First, the genetic structure of the north-Atlanticpopulations of common guillemots was investigated using 6microsatellites markers (Ibarguchi et al. 2000; Tirard et al.2002) and more than 600 birds sampled from 19 colonies fromthe north-east of the Atlantic Ocean (Fig. 1). Second, to assessthe reliability of the assignments tests for determining the ge-ographic origin of the oiled birds, we then attempted to assignthe individuals sampled in the 19 different candidate coloniesto their population of origin. We also used the same method toattempt to assign 48 ringed guillemots recovered following theoil spill. The test of the approach on birds of known origin wasneeded before considering any attempt on the large number ofsamples collected from oiled birds.

B. Cadiou et al.: Aquat. Living Resour. 17, 369–377 (2004) 371

Fig. 1. Distribution of common guillemot, Uria aalge, populations inthe north-east Atlantic and sampling locations (black circles) for thepopulation genetic study.

3 Results

3.1 Ring recoveries

Amongst the 266 ringed birds recorded during the oil spill,there were about three third of common guillemots, out ofwhich 184 had been ringed as chicks before fledging fromtheir colony of birth. The age and the geographic origin ofthese latter birds are thus known with certainty. Their ageranged from less than 1 to 18 years old (i.e. born from 1981 to1999) but 34% and 29% were young birds ringed in 1999 and1998 respectively. Amongst these individuals, 90 originatedfrom colonies located in western Scotland, 64 from south-eastern Irish colonies, 17 from Welsh colonies, 7 from northernScotland colonies and the remaining 6 from colonies borderingthe North Sea (4 from Britain and 2 from Helgoland, Germany;Fig. 2).

However these results are raw data which need to be con-sidered in relation to the location of the different colonieswhere ringing activities occurred in the north-east Atlantic andthe number of chicks ringed annually on these colonies. In-deed, ringing does not occur on all colonies and ringing ef-fort is not proportional to the abundance of the local breedingpopulations (Harris and Swann 2002). Thus, these figures donot reflect the real proportion of the birds originating from thedifferent geographic areas.

3.2 Biometrics

It was tested whether the wing length varied statisticallyaccording to the sex, the upperparts colour (brown or dark), thepresence of the cloacal bursa and the tip pattern of underwingcoverts (presence or absence of white tips). Results of the mul-tivariate analysis showed an effect of the four variables but not

Fig. 2. Geographic origin of the 184 common guillemots, Uria aalge,ringed as chicks and recovered in the Bay of Biscay during the Erikaoil spill (from Cadiou and Dehorter 2003).

of their interactions (SAS GLM procedure; sex: F1, 159 = 7.00,p < 0.01; upperparts colour: F1, 159 = 4.39, p < 0.05; cloa-cal bursa: F1, 159 = 15.56, p < 0.0001; underwing coverts:F1, 159 = 7.86, p < 0.01; N = 164 birds). Females had longerwings than males; older birds had longer wings than youngbirds and darkish birds have longer wings than brownish birds(Table 1). Differences also existed for head length, bill lengthand bill depth, older darkish males tending to be the largerindividuals (Table 1).

Mean wing lengths were then compared with data obtainedon breeding guillemots originating from different colonies ofthe north-east Atlantic. Only sexually mature birds (withoutcloacal bursa) of the Erika sample were taken into account be-cause immature birds are smaller, especially first year birds(Table 1). Both sexes were considered separately (Fig. 3). Adistinction was also made between birds with blackish orbrownish upperparts. Results showed that Erika’s brownishbirds had wing lengths situated between those of breedingbirds from the Celtic Sea and from western Scotland (Fig. 3).The blackish birds probably originated from Scotland as wellas potentially from colonies located in the Faeroe Islands andsouth-western Norway. Amongst these birds, there are also afew females with long wings which could come from muchmore northern colonies.

3.3 Genetic data

Overall, the level of population genetic differentiation, esti-mated using F-statistics, was very weak (average Fst = 0.004,p = 0.0108, 95% C.I. = 0.001−0.008). A plot of isolation by


Table 1. Biometrics of common guillemots collected during the Erika oil spill according to sex, upperparts colour, cloacal bursa and tip patternof underwing coverts. Mean ± standard deviation (mm), sample size in brackets, range in square brackets.

Sex Upper- Cloacal White tip Wing length Head length Bill length Bill depthparts bursa of

colour underwingcoverts

Female Brown Presence Presence 197.1 ± 5.5 (11) 108.9 ± 3.4 (32)46.2 ± 2.8 (32) 11.6 ± 0.7 (32)

[188-205] [100.0-114.9] [40.0-52.1] [10.3-13.6]Absence 200.1 ± 5.8 (19) 110.5 ± 2.8 (38)

46.9 ± 2.2 (38) 11.9 ± 0.5 (38)[192-212] [105.0-115.5] [41.5-50.1] [11.1-13.0]

Absence Presence 202.0 ± 7.9 (3) 110.9 ± 3.1 (9) 45.8 ± 2.9 (9) 12.4 ± 0.7 (9)[193-208] [107.0-115.0] [41.0-50.4] [11.3-13.4]

Absence 204.0 ± 4.9 (19) 110.2 ± 2.5 (34)46.8 ± 2.6 (34) 12.1 ± 0.6 (34)

[197-213] [105.0-116.0] [40.4-51.3] [10.9-13.6]Dark Presence Presence 202.6 ± 4.2 (5) 109.1 ± 3.9 (6) 45.0 ± 2.6 (6) 11.6 ± 0.9 (6)

[198-208] [105.0-115.5] [41.6-49.1] [11.0-13.3]Absence 202.9 ± 4.8 (13) 111.8 ± 2.3 (24)

47.8 ± 2.1 (24) 12.1 ± 0.6 (24)[196-211] [107.0-116.0] [41.8-51.2] [10.6-13.5]

Absence Presence 207.0 (1) 113.0 ± 0.0 (2) 47.6 ± 1.3 (2) 12.5 ± 0.8 (2)[113.0] [46.7-48.5] [11.9-13.1]

Absence 206.1 ± 6.5 (22) 111.9 ± 2.9 (32)47.4 ± 2.5 (32) 12.6 ± 0.5 (32)

[195-220] [107.0-118.5] [42.6-54.3] [11.9-13.7]Total – – 202.6 ± 6.2 (93) 46.8 ± 2.5 (177) 12.1 ± 0.7(177)

female 110.6 ± 3.0 (177)[188-220] [100.0-118.5] [40.0-54.3] [10.3-13.7]

Male Brown Presence Presence 195.7 ± 4.4 (17) 111.8 ± 2.7 (41)47.4 ± 2.4 (41) 11.9 ± 0.6 (41)

[188-203] [104.0-116.0] [40.8-51.5] [10.8-13.5]Absence 198.8 ± 4.9 (23) 113.8 ± 3.6 (40)

49.3 ± 2.6 (40) 12.4 ± 0.5 (40)[192-214] [107.0-121.0] [43.0-53.0] [11.1-13.6]

Absence Presence 201.0 (1) 114.5 ± 4.5 (6) 48.8 ± 3.6 (6) 12.2 ± 0.6 (6)[108.5-120.0] [42.6-53.8] [11.6-13.2]

Absence 202.2 ± 6.8 (13) 113.4 ± 3.7 (21)48.7 ± 2.6 (21) 13.0 ± 0.8 (21)

[192-214] [107.5-121.0] [45.0-53.5] [11.3-14.4]Dark Presence Presence 190.0 (1) 112.0 ± 0.9 (3) 50.2 ± 0.3 (3) 11.4 ± 0.5 (3)

[111.0-112.5] [49.9-50.5] [11.0-11.9]Absence 198.8 ± 3.1 (6) 113.3 ± 2.3 (10)

[195-203] 48.7 ± 2.2 (10) 12.2 ± 0.6 (10)[110.0-117.5] [45.5-52.6] [11.2-13.1]

Absence Presence 195.5 ± 6.4 (2) 112.5 ± 2.3 (3) 47.6 ± 1.9 (3) 12.5 ± 0.7 (3)[191-200] [110.0-114.5] [45.8-49.5] [11.7-12.9]

Absence 204.6 ± 5.8 (8) 114.6 ± 3.1 (20)[197-214] 49.7 ± 2.8 (20) 12.8 ± 0.6 (20)

[110.0-119.0] [44.9-54.5] [11.5-13.8]Total – – 199.2 ± 5.9 (71) 113.2 ± 3.3 (144) 48.6 ± 2.6 (144) 12.4 ± 0.8 (144)male

[188-214] [104.0-121.0] [40.8-54.5] [10.8-14.4]

Birds with cloacal bursa and white tip of underwing coverts are mainly first winter birds; birds without cloacal bursa and white tip of underwingcoverts are sexually mature birds and others are mainly immature birds (2−4 years old).


Fig. 3. Mean wing lengths (with range) of adults common guillemots, Uria aalge, originating from different colonies in the north-east Atlanticand of adults killed during the Erika oil spill (F = female; M = males; from Pethon 1967; Cramp 1985; Hope Jones 1984, 1988b; Camphuysen1989; Cadiou et al. 2003b; R. Barrett pers. comm.).

distance revealed a weak positive correlation between geneticand geographic distances (p = 0.047). This pattern of isola-tion by distance was much stronger when the samples from acolony of the western Atlantic (Witless Bay, Newfoundland,Canada) were included in the analyses (Riffaut et al., unpub-lished). Nevertheless, the low level of population genetic struc-ture prevented to determine reliably the origin of the oiledbirds: the results of the assignment tests showed that individ-uals from the breeding colonies could not be assigned reliablyto the colony in which they had been sampled (less than 6% ofthe individuals were assigned to the population in which theyhad been sampled). Further, the ringed birds recovered follow-ing the oil spill could not be assigned reliably to the coloniesof the region in which they had been ringed: a great majorityof the 48 ringed birds could be assigned to several areas oforigin despite the fact that 47 had been ringed in the Britishisles and one on the island of Helgoland, Germany (Fig. 4).These results were relatively disappointing given the wide areaof origin of the recovered birds, the polymorphism of the mi-crosatellite markers and the large spatial scale of the samplingof colonies (from Brittany in the south, to northern Norway inthe north-east).

4 Discussion

The common guillemots killed during the Erika oil spilloriginated from a large geographic area of the north-eastAtlantic. However, the great majority came from colonies lo-cated between western Scotland and the Celtic Sea. A lower

proportion originated from colonies alongside the North Sea,but also very likely from the Faeroe Islands or south-westernNorway. A much smaller proportion could have originatedfrom more northern colonies.

It was already known from ring recoveries that most ofthe young guillemots from Scotland spent their first winter inthe North Sea while those from the Celtic Sea winter in theBay of Biscay (Swann and Ramsey 1983; Harris and Swann2002). Analyses of ring recoveries during the Erika oil spillconfirmed this pattern of distribution (Cadiou and Dehorter2003). However, more data are still needed to estimate theproportion of birds from different origins which winter in thenorthern Bay of Biscay. Indeed, the pattern of dispersal ofcommon guillemots after the breeding season is complex andvaries according to their geographic origin, their age and prob-ably their sex (Brown 1985; Harris and Swann 2002). Also,prior to the Erika oil spill, there was no estimate of the numberof common guillemots at sea in the Bay of Biscay in winter.The first large scale aerial surveys started in the winter 2000-2001 (Bretagnolle et al. 2004). Therefore, it remains difficultto assess the potential impact of the mortality induced by theErika oil spill on the demography of the common guillemot inBritish and Irish colonies. On a short term time scale, no majordecrease in breeding numbers has been recorded in differentstudy colonies around the British Isles (Mavor et al. 2003).

The oil spill toll appears very high for common guillemots,with casualties ranging from 50 000 to 125 000 for recordedand estimated numbers respectively, including extrapolationsfor stranded corpses not counted on the coast and corpseslost at sea (see details in Cadiou et al. 2003a). However the


Fig. 4. Origin of the 48 oiled ringed common guillemots, Uria aalge,according to (1) the direct assignment method and (2) the “true” ori-gin of these oiled birds obtained from ring recovery data. For the di-rect assignment method. the 19 candidate populations were groupedby geographic region: Irish Sea, West Scotland and English Channel(Lambay, Great Saltee, Skomer and Cap Fréhel); East Scotland andShetland (Hermaness, Grunay, Noss, Foula, Sumburg Head, Fair Isleand Whinnyfold); East Europe (Stora Karlso and Helgoland) andNorth Europe (Hornøya, Røst, Sandoy Liraberg, Sandoy Honnin andSkuvoy Daer). These same geographic groupings are indicated for thetrue origins of the ringed birds.

demographic impact on the population could be far less obvi-ous especially because two main factors, the geographic areaof origin and the age structure of the birds, contribute to mit-igate the effect. Firstly, the potential area of origin held morethan one million breeding pairs in 1999, with 250 000 of themin the main area of origin, i.e. from the west of Scotlandto the Celtic Sea (Harris and Wanless 2004). The mean an-nual rate of increase of guillemot populations in Britain andIreland is 2% (Harris and Wanless 2004), the mean produc-tivity is 0.7 chick per breeding pair (Mavor et al. 2003), thesurvival rates is 46% during the first year of life, 75% duringthe second year, 88% during the third year and 95% afterwards

(Wernham et al. 1997; Harris et al. 2000) and breeding gen-erally does not occurred until birds are 5-6 years old (Harriset al. 1994). Taking into account these parameters, it can beestimated that the total number of birds from the main areaof origin was about 900 000 (500 000 breeders + 400 000 pre-breeders respectively, i.e. 175 000 juveniles born in 1999 and225 000 immatures born between 1995 and 1998), out ofwhich only an unknown proportion was effectively winteringin the north of the Bay of Biscay at the time of the Erika oilspill. Secondly, most of the common guillemots killed werenon-breeders that might have been recruited a few years later,with about one third of first winter birds, one third of oldersexually immature birds and one third of sexually mature birds(Cadiou et al. 2003b; see also Table 1). Moreover, the precisebreeding status of these latter birds remains unknown and theycould be potential first-time breeders or experienced breedingadults.

Although the use of the population genetic approach didnot help efficiently to determine the geographic origin of theoiled birds, this study provided some important informationabout the genetic structure and functioning of common guille-mot populations in the North Atlantic. From a genetic view-point, such results suggest that a management unit could in factbe the whole North Atlantic population (Friesen et al. 1997;Hedrick 2001). The weak genetic differentiation at this spatialscale revealed using highly polymorphic microsatellite mark-ers indeed shows that the colonies constitute a largely panmic-tic population. This idea supports previous results obtained us-ing mitochondrial markers (Moum et al. 1991; Friesen et al.1996; Friesen 1997; Moum and Arnason 2001). Moreover, theresults suggest that, despite the fact that a large number ofcommon guillemots were killed due to the Erika oil spill, lit-tle genetic variation is likely to have been lost. Despite a sup-posedly extreme rate of philopatry for this species (Tuck 1960;Hudson 1985; Harris and Wanless 1995; Harris et al. 1996a,b),the lack of strong population structure observed can be due toa historically recent expansion of the populations or can bethe direct result of intercolony dispersal. Dispersal can opposethe effect of the genetic drift and tends to genetically homoge-nize populations (Slatkin 1989). Intercolony dispersal has beenassumed for some common guillemot populations in whichcolony size increased too quickly to be explained solely by theintrinsic growth rate (Tuck 1960; Evans and Nettleship 1985;Harris and Wanless 2004). Resightings of ringed birds havealso suggested that auks, including the common guillemot, canmove and disperse between breeding colonies at large spatialscales (Halley and Harris 1993; Lyngs 1993; Kampp and Falk1998; R. Barrett, pers. comm.). If emigrants breed success-fully with local individuals, they may lead to sufficient geneflow to homogenize gene frequencies (Wright 1931; Hedrick1999). Significant intercolony dispersal could mean that com-mon guillemot populations have a strong capacity to recoverafter local disasters. The interplay between behavioural anddemographic processes is nevertheless important to consider(Boulinier and Lemel 1996). It has for instance been suggestedthat some type of population resilience may indeed be a con-sequence of dispersal (Cairns and Elliot 1987; Wiens 1996;Wiens et al. 1996) combined with increased access to breed-ing for first-time breeders due to the availability of vacant sites


(Croxall and Rothery 1991). In species in which individualsshow an attraction to conspecifics, a critical situation may nev-ertheless happen when local breeding groups decrease below acertain threshold size. Recruiting to an empty site is differentthan recruiting in an existing colony and an important ques-tion is the time it may take for a species to re-colonize an areafollowing local extinction.

The comparison of the results obtained in the context ofthe Erika oil spill with those following the Prestige oil spill,which occurred three years later than the wreck of Erika inthe south-western part of the Bay of Biscay, should be veryinteresting. Indeed, once again, the common guillemot was themost impacted species (Camphuysen et al. 2002; García et al.2003).

Complementary analyses based on stable isotopes couldalso be used to determine the origin of birds (e.g. Hobson1999; Chérel et al. 2000; Lott et al. 2003). Samples of feath-ers, especially from common guillemots and divers Gavia spp.,have been collected and stored during the Erika oil spill in or-der to investigate the usefulness of this tool on a conservationviewpoint.

5 Conclusion

The results obtained underline that for a full assessment ofthe ecological impact of an oil spill on seabird populations, itis necessary to combine information on the dynamics of thedistribution of seabirds at sea with knowledge of the differentprocesses involved in the dynamics of the subdivided breed-ing populations. Investigations using model simulations underdifferent scenarios should help integrating the available infor-mation for seabirds sensitive to oil pollution (Ford et al. 1982;Seip et al. 1991). In the case of oil spills, which occur often ata time when birds are dispersed away from the colonies, notonly the age and sex structure of the population must be con-sidered, but also the potential mixing of different subgroupswhich will vary with the time of the year. Such a modellingapproach should allow one to use the available informationto assess the potential consequences of a specific oil spill ondifferent seabird populations, but also more generally to deter-mine the scale and type of data needed to assess the potentialimpact on the breeding populations.

During an oil spill, three points relevant to birds should becarefully investigated in order to allow an ecological impactassessment on a short, medium or long term scale. First, allavailable data should be centralized and analysed to obtain thetotal number of living or dead oiled birds collected on the coastand, as far as possible, to estimate the global mortality. Sec-ond, studies should be conducted to determine the geographicorigins of the birds. Third, studies should also be conductedto determine the population structure of the species impacted(sex-ratio, age ratio and sexual maturity). Indeed, as seabirdsare long-lived species exhibiting deferred maturity, the demo-graphic impact of massive mortality would not be the same ifthe majority of the victims are young non-breeding birds orbreeding adults (Croxall and Rothery 1991; Russell 1999).

Mass mortality during oil spill offer opportunity to obtainlarge samples of seabirds for different studies and this biolog-ical material should not be wasted. Indeed, due to technical,

legal or ethical reasons it is impossible to take such samplesunder normal circumstances.

Experience from the Erika oil spill has highlighted prob-lems and gaps and the crucial need of more coordination be-tween administrative authorities, scientists and rehabilitationcentres receiving oiled birds (Cadiou et al. 2003a,b). Pre-planning is essential to allow an immediate and relevant re-sponse and to avoid the lack of preparedness as during theErika and Prestige oil spills (Cadiou et al. 2003a; Heubecket al. 2003).

Complementary studies based on classical methods (ringrecoveries and biometrics) or more recent techniques (geneticand stable isotopes) appear essential to improve the determi-nation of the geographic origins of birds of different species.Seabird corpses stranded on the coast in winter should beconsidered as a valuable biological material in order to in-crease data collection (Camphuysen and van Franeker 1992).All these studies would allow giving more precise answers andwould contribute to better oil spill impact assessment in thefuture.

Acknowledgements. We thank all the people and organizations in-volved in fieldwork during the Erika oil spill. We are most gratefulto Tycho Anker Nilssen, Vidar Bakken, Rob Barrett, Tim Birkhead,Olivier Duriez, Morten Frederiksen, Bob Furness, Jeff Graves, PaulHarvey, Martin Heubeck, Ommo Hueppop, David Jardine, OscarMerne, Steve Newton, Bergur Olsen, Olof Olsson, Henrik Osterblom,Aevar Petersen, Derek Shaw, Susan Waugh and Chris Wernham forhelping with the collection of DNA samples of birds in breedingcolonies. Vicki Friesen, Elodie Chapuis, Humberto Cabrera, MelanieDouadi, Julien Gasparini, Fabrice Helfelstein, Thomas Regembal andMurielle Richard kindly helped at different stages of the geneticwork as well as Ibtissem Chaherdine, Nathalie Delliou, Régis Gallais,Séverine Gibet, Gabrielle Guillopé, Yves Le Bail, François Le Poul,Catherine Orain and Fred Touzalin for the biometric work. Discus-sions with Torkild Tveraa and Per Fauchald were influential at earlystages of the population genetic work. We are also grateful to theanonymous referee who contributed to improve the manuscript. Wethank the various authorities, in France and abroad, from which weobtained permits to collect and gather samples. The studies benefitedfrom support by the French Ministry of Ecology (Programmes “SuiviErika” and “Liteau-Erika”), the French Embassy of Canada (FondsFrance-Canada pour la Recherche and Chateaubriand Post-doctoralaward), the French Polar Institute (IPEV, Programme No. 333), theC.N.R.S. (France), NSERC, Canada and the Queen’s UniversityInternational Visitor Program (Canada).

References

Anker-Nilssen T., Hope Jones P., Røstad O.W., 1988, Age, sex andorigins of auks (Alcidae) killed in the Skagerrak oiling incidentof January 1981. Seabird 11, 28-46.

Bretagnolle V., Certain G., Houte S., Métais M., 2004, Distributionmaps and minimum abundance estimates for wintering auksin the Bay of Biscay, based on aerial surveys. Aquat. LivingResour. 17, 353.

Boulinier T., Lemel J.-Y., 1996, Spatial and temporal variations offactors affecting breeding habitat quality in colonial birds: someconsequences for dispersal and habitat selection. Acta Oecol. 17,531-552.


Brown R.G.B., 1985, The atlantic alcidae at sea. In: Nettleship D.N.,Birdhead T.R. (Eds.), The Atlantic Alcidae. Academic Press,London, pp. 383-426.

Cadiou B., 2002, Oiseaux marins nicheurs de Bretagne. Les cahiersnaturalistes de Bretagne No 4. Conseil Régional de Bretagne,Éditions Biotope, Mèze.

Cadiou B., Dehorter O., 2003, Marée noire de l’Erika –Contribution à l’étude de l’impact sur l’avifaune. Analyse desreprises/contrôles de bagues. Rapport Bretagne Vivante-SEPNB,CRBPO, DIREN Bretagne.

Cadiou B., Chenesseau D., Joslain H., 2003a, Marée noire de l’Erika– Contribution à l’étude de l’impact sur l’avifaune. Bilan nationaldes échouages et de la mortalité des oiseaux (BNEMO). RapportBretagne Vivante-SEPNB, LPO Loire-Atlantique, Observatoiredes marées noires, DIREN Bretagne.

Cadiou B., Cam E., Fortin M., Monnat J.-Y., Gélinaud G., CabelguenJ., Le Roch A., 2003b, Impact de la marée noire de l’Erika surles oiseaux marins migrateurs : détermination de l’origine et dela structure des populations par la biométrie. Rapport BretagneVivante-SEPNB, DIREN Bretagne.

Cairns D.K., Elliot R.D., 1987, Oil spill impact assessment forseabirds: the role of refugia and growth centers. Biol. Cons. 40,1-9.

Camphuysen C.J., 1989, Biometrics of auks at Jan Mayen. Seabird12, 7-10.

Camphuysen C.J., 1995, Leeftijdsbepaling van Zeekoet Uria aalgeen Alk Alca torda in de hand. Sula 9, 1-22.

Camphuysen C.J., van Franeker J.A., 1992, The value of beachedbird surveys in monitoring marine oil pollution. Technisch rap-port Vogelbescherming 10, Vogelbescherming Nederland, Zeist.

Camphuysen C.J., Heubeck M., Cox S.L., Bao R., Humple D.,Abraham C., Sandoval A., 2002, The Prestige oil spill in Spain.Atlantic Seabirds 4, 131-140.

Chérel Y., Hobson K.A., Weimerskirch H., 2000, Using stable-isotope analysis of feathers to distinguish moulting and breedingorigins of seabirds. Oecologia 122, 155-162.

Cornuet J.M., Piry S., Luikart G., Estoup A., Solignac M., 1999, Newmethods employing multilocus genotypes to select or excludepopulations as origins of individuals. Genetics 153, 1989-2000.

Cramp S. (Ed.), 1985, The Birds of the Western Palearctic, Vol. IV.Oxford University Press, Oxford, London, New York.

Croxall J.P., Rothery P., 1991, Population regulation of seabirds: im-plications of their demography for conservation. In: Perrins C.M.,Lebreton J.-D., Hirons G.J.M. (Eds.), Bird population studies:relevance to conservation and management. Oxford UniversityPress, Oxford, pp. 272-296.

Edwards S.V., Siva M.C., Burg T., Friesen V., Warheit K.I., 2001,Molecular genetic markers in the analysis of seabird bycatch pop-ulations. In Proceedings – Seabird bycatch: Trends, roadblocksand solutions. University of Alaska Sea Grant, pp. 115-140.

Ford R.G., Wiens J.A., Heineman D., Hunt G.L., 1982, Modelling thesensitivity of colonially breeding marine birds to oil spills: guille-mot and kittiwake populations on the Pribilof Islands, Bering Sea.J. Appl. Ecol. 19, 1-31.

Friesen V.L., Montevecchi W.A., Baker A.J., Barrett R.T., DavidsonW.S., 1996, Population differentiation and evolution in the com-mon guillemot Uria aalge. Mol. Ecol. 5, 793-805.

Friesen V.L., 1997, Population genetics and the spatial scale of con-servation of colonial waterbirds. Colon. Waterbird 20, 353-368.

García L., Viada C., Moreno-Opo R., Carboneras C., Alcade A.,González F., 2003, Impacto de la marea negra del Prestige so-bre las aves marinas. SEO/Birdlife, Madrid.

Halley D.J., Harris M.P., 1993, Intercolony movement and behaviourof immature Guillemots Uria aalge. Ibis 135, 264-270.

Harris M.P., Halley D.J., Swann R.L., 1994, Age of first breeding inCommon Murres. Auk 111, 207-209.

Harris M.P., Halley D.J., Wanless S., 1996a, Philopatry in theCommon Guillemot Uria aalge. Bird Study. 43, 134-137.

Harris M.P., Wanless S., 1995, Survival and non-breeding of adultCommon Guillemots Uria aalge. Ibis 137, 192-197.

Harris M.P., Wanless S., Barton T.R., 1996b, Site use and fidelity inthe Common Guillemot Uria aalge. Ibis 138, 399-404.

Harris M.P., Swann B., 2002, Common guillemot. In: Wernham C.V.,Toms M.P., Marchant J.H., Clark J.A., Siriwardena G.M., BaillieS.R. (Eds.), The migration atlas: movements of birds of Britainand Ireland. T. & A.D. Poyser, London, pp. 397-400.

Harris M.P., Wanless S., 2004, Common guillemots Uria aalge. In:Mitchell, I., Newton, S., Ratcliffe, R., Dunn, T.E. (Eds.), Seabirdpopulations of Britain and Ireland. T. & A.D. Poyser, London,pp. 350-363.

Harris M.P., Wanless S., Rothery P., Swann R.L., Jardine D., 2000,Survival of adults common guillemot Uria aalge at three Scottishcolonies. Bird Study 47, 1-7.

Hedrick P.W., 1999, Population Genetics. Jones and Bartlett, Boston,USA. 2nd edition.

Hedrick P.W., 2001, Conservation genetics: where are we now?Trends Ecol. Evol. 16, 629-636.

Heubeck M., Camphuysen K., Bao Casal R., Humple D., SandovalRey A., Cadiou B., Bräger S., Thomas T., 2003, Assessing theimpact of major oil spill on seabird populations. Mar. Pollut. Bull.46, 900-902.

Hobson K.A., 1999, Tracing origins and migration of wildlife usingstable isotopes: a review. Oecologia 120, 314-326.

Hope Jones P., 1984, Skins of guillemots Uria aalge and razorbillsAlca torda examinated at Cascais, Portugal, in May 1982. Mem.Mus. Mar. Portugal 3 (27), 1-10.

Hope Jones P., 1988, The European cline in wing-length of guillemotsUria aalge. Seabird 11, 19-21.

Hope Jones P., Blake B.F., Anker-Nilssen T., Rostad O.W., 1982a,The examination of birds killed in oilspills and other incidents -a manual of suggested procedures. Nature Conservancy Council,Aberdeen.

Hope Jones P., Monnat J.-Y., Harris M.P., 1982b, Origins, age and sexof auks (Alcidae) billed in the ‘Amoco Cadiz’ oiling incident inBrittany, March 1978. Seabird Rep. 6, 122-130.

Hudson P.J., 1985, Population Parameters for the Atlantic Alcidae.In: Nettleship D.N., Birdhead T.R. (Eds.), The Atlantic Alcidae.Academic Press, London, pp. 233-261.

Ibarguchi G., Birt T.P., Warheit K.I., Boag P.T., Friesen V.L., 2000,Microsatellite loci from common and thick-billed murres, Uriaaalge and Uria lomvia. Mol. Ecol. 9, 638-639.

Kampp K., Falk K., 1998, A long distance colony shift by a thick-billed murre. Colon. Waterbird 21, 91-93.

Lott C.A., Meehan T.D., Heath J.A., 2003, Estimating the latitudinalorigins of migratory birds using hydrogen and sulphur stable iso-topes in feathers: influence of marine prey base. Oecologia 134,505-510.

Lyngs P., 1993, Colony interchange in Baltic Guillemots Uria aalge.Dansk Orn. Foren. Tidsskr 87, 247-250.

Mavor R.A., Parsons M., Heubeck M., Pickerell G., Schmitt S., 2003,Seabird numbers and breeding success in Britain and Ireland,2002. Joint Nature Conservation Committee, Peterborough, UKNature Conservation, No. 27.

Møller A.P., Erritzøe J., 2001, Dispersal, vaccination and regressionof immune defence organs. Ecol. Lett. 4, 484-490.


Moum T., Erikstad K.E., Bjorkild E., 1991, Restriction fragment anal-ysis of mitochondrial DNA in Common Murres, Uria aalge, fromfour Norwegian seabird colonies. Can. J. Zool. 69, 1577-1584.

Moum T., Arnason E., 2001, Genetic diversity and population his-tory of two related seabird species based on mitochondrial DNAcontrol region sequences. Mol. Ecol. 10, 2463-2478.

Nettleship D.N., Evans P.G.H., 1985, Distribution and Status of theAtlantic Alcidae. In: Nettleship, D.N., Birdhead T.R. (Eds.), TheAtlantic Alcidae. Academic Press, London, pp. 53-154.

Piatt J.F., Lensink C.J., 1998, Exxon Valdez bird toll. Nature 242,865-866.

Pethon P., 1967, The systematic position of the Norwegian commonmurre (Uria aalge) and puffin (Fratercula arctica). Nytt Mag.Zool. 14, 84-95.

Russell R.W., 1999, Comparative demography and life history tacticsof seabirds: implications for conservation and marine monitoring.Am. Fish. Soc. Symp. 23, 51-76.

Seip K.L., Sandersen E., Mehlum F., Ryssdal J., 1991, Damages toseabirds from oil spills: comparing simulation results and vulner-ability indexes. Ecol. Model. 53, 39-60.

Slatkin M., 1989, Detecting small amounts of gene flow fromphylogenies of alleles. Genetics 121, 609-612.

Swann R.L., Ramsay D.K., 1983, Movements from and age ofreturn to an expanding Scottish guillemot colony. Bird Study 30,207-214.

Tirard C., Helfenstein F., Danchin E., 2002, Polymorphic microsatel-lites in the black-legged kittiwake, Rissa tridactyla. Mol. Ecol.Notes 2, 431-433.

Tuck L.M., 1960, Les Marmettes. Duhamel, Ottawa, Canada.Weir D., Kitchener A., McGowan R., Kinder A., Zonfrillo B., 1997,

Origins, population structure, pathology and diet of samples ofdiver and auk casualties of the Sea Empress oil spill. CCW SeaEmpress Contract Report.

Wernham C.V., Peach W.J., Browne S.J., 1997, Survival rates ofrehabilitated guillemots. BTO Res. Rep. No. 186.

Wiens J.A., 1996, Oil, seabirds and science. BioScience 46, 587-597.Wiens J.A., Crist T.O., Day R.H., Murphy R.H., Hayward G.D., 1996,

Effects of the Exxon Valdez oil spill on marine bird communitiesin Prince William Sound, Alaska. Ecol. Appl. 6, 828-841.

Wright S., 1931, Evolution in Mendelian populations. Genetics 16,111-123.

ecological impact of the “erika” oil spill: determination of the

Documents